Datasets:

sayakpaul
/

coco-30-val-2014

Modalities:

Formats:

Size:

Libraries:

Dataset card Viewer Files Files and versions Community

sayakpaul HF staff commited on Jan 31

Commit

aaa5150

•

1 Parent(s): bc5bd10

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -14,6 +14,6 @@ dataset_info:
 ---
 # Dataset Card for "coco-30-val-2014"
-This is 30k randomly sampled image-captioned pairs from the COCO 2014 `val` split. This is useful for image generation benchmarks (FID, CLIPScore, etc.).
 Refer to the gist to know how the dataset was created: https://gist.github.com/sayakpaul/0c4435a1df6eb6193f824f9198cabaa5.

 ---
 # Dataset Card for "coco-30-val-2014"
+This is 30k randomly sampled image-captioned pairs from the [COCO](https://cocodataset.org/) 2014 `val` split. This is useful for image generation benchmarks (FID, CLIPScore, etc.).
 Refer to the gist to know how the dataset was created: https://gist.github.com/sayakpaul/0c4435a1df6eb6193f824f9198cabaa5.